PDF: GWPCARE3
(2018)Objective
To evaluate the efficacy and safety of two doses of cannabidiol (10 mg/kg/day and 20 mg/kg/day) as adjunctive therapy for drop seizures in patients with the Lennox-Gastaut syndrome
Study Summary
• Cannabidiol 10 mg/kg/day reduced drop-seizure frequency by 37.2% vs 17.2% with placebo (median difference 19.2 pp; P=0.002)
• ≥50% reduction in drop seizures: 39% (20 mg), 36% (10 mg) vs 14% placebo (both P<0.005)
• Total seizure frequency reduced by 38.4% (20 mg) and 36.4% (10 mg) vs 18.5% placebo
• Caregiver Global Impression of Change improved in 57% (20 mg), 66% (10 mg) vs 44% placebo
• Elevated liver aminotransferases (>3× ULN) occurred in 9% of cannabidiol patients, particularly with concomitant valproate
Intervention
Cannabidiol oral solution at 20 mg/kg/day or 10 mg/kg/day (escalated over 2 weeks, then 12 weeks maintenance) added to stable antiepileptic drug regimen vs matching placebo for 14 weeks
Inclusion Criteria
• Ages 2–55 years
• Lennox-Gastaut syndrome with EEG showing slow spike-and-wave (<3.0 Hz)
• ≥2 types of generalized seizures including drop seizures for ≥6 months
• ≥2 drop seizures per week during baseline
• On 1–4 antiepileptic drugs (stable for ≥4 weeks)
• Ketogenic diet and VNS stable for ≥4 weeks
Study Design
Arms: Array
Patients per Arm: 20-mg CBD: 76; 10-mg CBD: 73; Placebo: 76
Outcome
• 20-mg vs placebo: median difference 21.6 pp (95% CI 6.7–34.8; P=0.005)
• 10-mg vs placebo: median difference 19.2 pp (95% CI 7.7–31.2; P=0.002)
• ≥50% drop-seizure reduction: 39% (20 mg; OR 3.85, P<0.001), 36% (10 mg; OR 3.27, P=0.003) vs 14% placebo
• Total seizure reduction: 38.4% (20 mg) vs 36.4% (10 mg) vs 18.5% placebo
• Adverse events: somnolence, decreased appetite, diarrhea, elevated LFTs (9%)
Bottom Line
Both doses of cannabidiol (10 and 20 mg/kg/day) significantly reduced drop-seizure frequency compared to placebo in patients with Lennox-Gastaut syndrome. The 20-mg dose reduced drop seizures by 42% and the 10-mg dose by 37%, compared to 17% with placebo. Both doses also improved total seizure frequency and caregiver-rated overall condition. Elevated liver aminotransferases occurred in 9% of CBD patients, primarily those on concomitant valproate. The 10-mg dose had a more favorable adverse-event profile.
Major Points
- Phase 3, three-arm trial testing two doses of cannabidiol vs placebo for drop seizures in Lennox-Gastaut syndrome
- Both CBD doses significantly reduced drop-seizure frequency: 20 mg (−41.9%, P=0.005) and 10 mg (−37.2%, P=0.002) vs placebo (−17.2%)
- Dose-response for drop seizures was modest — 10-mg dose was nearly as effective as 20-mg dose with fewer side effects
- 39% (20 mg) and 36% (10 mg) of patients achieved ≥50% drop-seizure reduction vs 14% placebo
- No patients were completely free of drop seizures during the entire treatment period, though 5% (20 mg) and 4% (10 mg) were free during maintenance
- Total seizure frequency was also significantly reduced with both doses
- Caregiver Global Impression of Change improved in 57% (20 mg) and 66% (10 mg) vs 44% placebo
- Elevated liver aminotransferases (>3× ULN) occurred in 9% of CBD patients, most commonly in those receiving concomitant valproate
Study Design
- Study Type
- Phase 3, multicenter, randomized, double-blind, placebo-controlled trial
- Randomization
- Yes
- Blinding
- Double-blind
- Sample Size
- 225
- Follow-up
- 14 weeks treatment (2-week escalation + 12-week maintenance) plus taper and 4-week safety follow-up
- Centers
- 30
- Countries
- United States, Spain, United Kingdom, France
Primary Outcome
Definition: Percentage change from baseline in monthly drop-seizure frequency
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| −17.2% | 20 mg: −41.9%; 10 mg: −37.2% | - (20 mg vs placebo: 6.7-34.8; 10 mg vs placebo: 7.7-31.2 (median difference)) | 20 mg: P=0.005; 10 mg: P=0.002 |
Limitations & Criticisms
- Short treatment duration (14 weeks); long-term efficacy and safety unknown
- No patients were completely seizure-free during the entire treatment period
- Elevated aminotransferases in 9% of CBD patients, primarily on valproate — safety concern for long-term use
- Functional unblinding possible due to recognizable side effects
- Concomitant clobazam use may have contributed to apparent efficacy via pharmacokinetic interaction
- 2:2:1:1 randomization means fewer patients in placebo groups, reducing statistical power for safety comparisons
- Industry funded (GW Pharmaceuticals)
Citation
N Engl J Med 2018;378:1888-1897